Isolation of Nocuolin A and Synthesis of New Oxadiazine Derivatives. Design, Synthesis, Molecular Docking, Apoptotic Evaluation, and Cathepsin B Inhibition

Nocuolin A (1), an oxadiazine, was isolated from the cyanobacterium Nostoc sp. Its chemical structure was elucidated using NMR and mass spectroscopic data. From this compound, two new oxadiazines, 3-[(6R)-5,6-dihydro-4,6-dipentyl-2H-1,2,3-oxadiazin-2-yl]-3-oxopropyl acetate (2) and 4-{3-[(6R)-5,6-dihydro-4,6-dipentyl-2H-1,2,3-oxadiazin-2-yl]-3-oxopropoxy}-4-oxobutanoic acid (3), were synthesised. The chemical structures of these two compounds were elucidated by a combination of NMR and MS analysis. Compound 3 showed cytotoxicity against the ACHN (0.73 ± 0.10 μM) and Hepa-1c1c7 (0.91 ± 0.08 μM) tumour cell lines. Similarly, compound 3 significantly decreased cathepsin B activity in ACHN and Hepa-1c1c7 tumour cell lines at concentrations of 1.52 ± 0.13 nM and 1.76 ± 0.24 nM, respectively. In addition, compound 3 showed no in vivo toxicity in a murine model treated with a dose of 4 mg/kg body weight.


Introduction
Protein degradation, in both intracellular and extracellular spaces, is essential for the physiological equilibrium in healthy and diseased cells. It must, therefore, be strictly monitored [1]. Chronic diseases, especially cancer, are characterised by the dysregulation of proteolysis, which contributes to the progression of the disease [2]. Within this protein catabolism, lysosomal proteases have been identified. Among these proteases, several cysteine cathepsins are overexpressed in tumours [3].
Cathepsins are a large family of proteases composed of at least twelve different molecules that catalyse the hydrolysis of different proteins. There are three main types of enzymes depending on the amino acid of their active site, which can include a cysteine, aspartate, or a serine. Each of them has a unique structure, catalytic mechanism, and substrate specificity [4]. One of the degradation processes includes the macroautophagy pathway, where dysfunctional organelles are enclosed by a phagophore or isolation membrane (IM) that is later expanded for the creation of an autophagosome, a double-membrane vacuole that engulfs cellular components. Subsequently, the autophagosome fuses with lysosomes, where the lysosomal cathepsins, i.e., cathepsin B, D, L, etc., play a key role in allowing the normal function of the autophagosome [5]. This fusion gives rise to the creation of autolysosomes, which are essential structures for organelle degradation [6]. Additionally, cathepsin proteolytic activity has been directly linked with the determination of tumour

Synthesis of Analogues of the Natural Compound
The analogues doxadiazine 2 (compound 2) and doxadiazine 3 (compound 3) were synthesised from compound 1 by esterification reactions. In this sense, the reaction of compound 1 with acetic anhydride in trimethylamine yielded 88% of compound 2 (colourless oil). On the other hand, to obtain compound 3, compound 1 was esterified with succinic anhydride and trimethylamine, obtaining it with a 60% yield and a colourless oil (Scheme 1). Scheme 1. Chemical synthesis of analogues 2 and 3 from compound 1.
The elucidation of the structures of derivatives 2 and 3 was carried out by comparison with the 1D and 2D spectra in CDCl3 of compound 1. The complete 1D NMR ( 1 H and 13 C) assignment of compound 1 was possible only after recording several 1D and 2D ( 1 H-1 H COSY, 1 H-13 C HSQC, 1 H-13 C HSQC-TOCSY, 1 H-13 C HMBC and 1 H-15 N HMBC) in CDCl3 at 500 MHz ( Figures S9-S16). Subsequently, the NMR spectral data of both derivatives 2 and 3 were compared with compound 1's data (see experimental section).
Compound 2, which was obtained by synthesis from compound 1 as a colourless oil, gave the molecular formula of C18H32N2O4, obtained by HRESIMS at m/z 363.2444 (calcd.

Synthesis of Analogues of the Natural Compound
The analogues doxadiazine 2 (compound 2) and doxadiazine 3 (compound 3) were synthesised from compound 1 by esterification reactions. In this sense, the reaction of compound 1 with acetic anhydride in trimethylamine yielded 88% of compound 2 (colourless oil). On the other hand, to obtain compound 3, compound 1 was esterified with succinic anhydride and trimethylamine, obtaining it with a 60% yield and a colourless oil (Scheme 1).
Compound 1 (Figure 1) was patented by the company VALORALIA I MÁS D, SL [23]. However, in the same month of filing the acceptance of the patent process, a patent was made public in the Czech Republic, describing exactly the same compound, which they named Nocuolin A. Subsequently, the same research group published an article with the structural characterisation and antitumour activity of compound 1 [24]. On the other hand, the compound Nocuolin A has also been reported in the species Nodularia sp. [25]. Despite this, the company VALORALIA I MÁS D, SL has ownership of the exploitation of the uses of the invention in European, Asian, and American countries.

Synthesis of Analogues of the Natural Compound
The analogues doxadiazine 2 (compound 2) and doxadiazine 3 (compound 3) were synthesised from compound 1 by esterification reactions. In this sense, the reaction of compound 1 with acetic anhydride in trimethylamine yielded 88% of compound 2 (colourless oil). On the other hand, to obtain compound 3, compound 1 was esterified with succinic anhydride and trimethylamine, obtaining it with a 60% yield and a colourless oil (Scheme 1). Scheme 1. Chemical synthesis of analogues 2 and 3 from compound 1.
The elucidation of the structures of derivatives 2 and 3 was carried out by comparison with the 1D and 2D spectra in CDCl3 of compound 1. The complete 1D NMR ( 1 H and 13 C) assignment of compound 1 was possible only after recording several 1D and 2D ( 1 H-1 H COSY, 1 H-13 C HSQC, 1 H-13 C HSQC-TOCSY, 1 H-13 C HMBC and 1 H-15 N HMBC) in CDCl3 at 500 MHz ( Figures S9-S16). Subsequently, the NMR spectral data of both derivatives 2 and 3 were compared with compound 1's data (see experimental section).
Compound 2, which was obtained by synthesis from compound 1 as a colourless oil, gave the molecular formula of C18H32N2O4, obtained by HRESIMS at m/z 363.2444 (calcd. The elucidation of the structures of derivatives 2 and 3 was carried out by comparison with the 1D and 2D spectra in CDCl 3 of compound 1. The complete 1D NMR ( 1 H and 13 C) assignment of compound 1 was possible only after recording several 1D and 2D ( 1 H-1 H COSY, 1 H-13 C HSQC, 1 H-13 C HSQC-TOCSY, 1 H-13 C HMBC and 1 H-15 N HMBC) in CDCl 3 at 500 MHz ( Figures S9-S16). Subsequently, the NMR spectral data of both derivatives 2 and 3 were compared with compound 1's data (see experimental section).
Compound 2, which was obtained by synthesis from compound 1 as a colourless oil, gave the molecular formula of C 18 Figure S17). The analysis of the 1D NMR spectra ( 1 H, 13 C, and DEPT-135) (Figures S18-S20) suggested that the only difference between the 1D NMR spectra of compound 2 and the 1D NMR spectra of compound 1 was the presence of one acetate group.
Compounds 2 and 3, named Doxadiazine 2 and Doxadiazine 3, respectively, were patented by the VALORALIA I MÁS D, SL company, having ownership of the uses of the invention in countries of Europe, Asia, and America [26].

Inhibition of Cathepsin B Activity
For the determination of the cathepsin B's inhibition activity, its superficial expression in two tumour cell lines (ACHN and Hepa-1c1c7 cells) and two non-tumour cell lines (TKPTS and FL83B cells) was evaluated by flow cytometry. CA 074 is an inhibitor of cathepsin B and was used as a positive control with Ki values of 2.39 ± 0.12 and 3.08 ± 0.19 nM for the non-tumour cell lines TKPTS and FL83B. Likewise, the positive control showed Ki values of 5.62 ± 0.73 and 5.99 ± 0.64 nM for the tumour cell lines ACHN and Hepa-1c1c7 ( Figure 3A).
Compound 1 decreased the amount of cathepsin B on tumour cell lines similarly to the positive control (p = 0.5374), but with lower Ki, 3.06 ± 0.13 nM (ACHN cells) and 3.37 ± 0.24 nM (Hepa-1c1c7 cells). However, compound 1 also decreased the amount of cathepsin B in the FL83B non-tumour cell line (p = 0.8511), with similar results to the positive control, obtaining a Ki of 2.61 ± 0.02 nM. However, this last effect was smaller on the nontumour cell line TKPTS at a Ki of 2.13 ± 0.04 nM (p < 0.001, compared to the positive control) Compounds 1 and 3 did not show significant cytotoxicity compared to Actinomycin in the TKPTS non-tumour cell lines (CC 50 of 1.10 ± 0.05 and 1.62 ± 0.06 µM, respectively) and FL83B (CC 50 of 2.36 ± 0.10 and 2.80 ± 0.09 µM, respectively). Concerning their cytotoxicity on tumour cell lines, compounds 1 and 3 showed CC 50 values of 0.76 ± 0.09 and 0.73 ± 0.10 µM (ACHN cells) and CC 50 of 1.25 ± 0.06 and 0.91 ± 0.08 µM (Hepa-1c1c7 cells).
On the other hand, compound 2 presented higher cytotoxicity than compounds 1 and 3 on tumour cell lines ACHN and Hepa-1c1c7 with CC 50 values of 0.22 ± 0.03 and 0.63 ± 0.05 µM, respectively. However, compound 2 also showed significant cytotoxicity against the non-tumour cell lines TKPTS and FL83B with CC 50 values of 1.06 ± 0.04 and 1.94 ± 0.08 µM, respectively ( Figure 2).

Inhibition of Cathepsin B Activity
For the determination of the cathepsin B's inhibition activity, its superficial expression in two tumour cell lines (ACHN and Hepa-1c1c7 cells) and two non-tumour cell lines (TKPTS and FL83B cells) was evaluated by flow cytometry. CA 074 is an inhibitor of cathepsin B and was used as a positive control with K i values of 2.39 ± 0.12 and 3.08 ± 0.19 nM for the non-tumour cell lines TKPTS and FL83B. Likewise, the positive control showed K i values of 5.62 ± 0.73 and 5.99 ± 0.64 nM for the tumour cell lines ACHN and Hepa-1c1c7 ( Figure 3A). Cathepsin B is a 30-kDa, bilobal protein with an active site and substrate-binding cleft located in the interface of a left (L-) and a right (R-) domain with three and two binding sites located in the loops, categorised as S3, S1, and S2', and S1' and S2, respectively. Access of the substrate into the active site is controlled by the occluding loop, which consists of an 18-residue-long insertion (Pro 107-Asp 124) [27].
Based on the structure of this enzyme, it is known that the catalytic triad of cathepsin B is formed by a cysteine, a histidine, and an aspartic acid. The Cys29 and His199 residues interact, catalysing the cleavage of the peptide bond. Therefore, cathepsin B inhibitors must have an electrophilic character capable of reacting with the thiol group of the Cys29 residue [17]. Examples include synthetic 'warheads' such as aldehydes, disulphides, vinyl sulfones, and halomethyl ketones [28]. In our work, we used the compound CA 074 as a positive control, whose mechanism of action is performed through its binding to the S' subsites in a similar direction as the substrate [17].
All the compounds (one isolated and two synthesised) share the essential premise of maintaining a main oxadiazine scaffold, which is structurally similar to thiadiazols. In this sense, a cleavage of the N-O bond of oxadiazine by the thiolate group of cathepsin B could produce the opening of the ring and proteolytic enzymatic inhibition [17].
However, in the case of compound 2, to improve the membrane permeability, an ester-type derivative was introduced. Nevertheless, the activity results obtained were worse (loss of selectivity) compared to compound 1 and the positive control. These data agree Compound 1 decreased the amount of cathepsin B on tumour cell lines similarly to the positive control (p = 0.5374), but with lower K i , 3.06 ± 0.13 nM (ACHN cells) and 3.37 ± 0.24 nM (Hepa-1c1c7 cells). However, compound 1 also decreased the amount of cathepsin B in the FL83B non-tumour cell line (p = 0.8511), with similar results to the positive control, obtaining a K i of 2.61 ± 0.02 nM. However, this last effect was smaller on the non-tumour cell line TKPTS at a K i of 2.13 ± 0.04 nM (p < 0.001, compared to the positive control) ( Figure 3A). Compounds 2 and 3 showed a higher decrease in the amount of cathepsin B from tumour cell lines than the positive control (p < 0.001). However, compound 3 showed this effect at a lower K i (K i = 1.52 ± 0.13 nM and K i = 1.76 ± 0.24 nM, in ACHN cells and Hepa-1c1c7 cells, respectively) than the positive control. On the other hand, compound 2 showed a similar effect (K i = 5.67 ± 0.68 nM, ACHN cells and K i = 6.59 ± 0.65 nM, Hepa-1c1c7 cells) to the positive control ( Figure 3A).
Although compound 2 showed the potential of decreasing the amount of cathepsin B on tumour cell lines, it was observed that it also significantly affected non-tumour cell lines compared to the positive control (p < 0.001), reporting K i of 0.49 ± 0.05 nM (TKPTS cells) and 0.66 ± 0.09 nM (FL83B cells). In the case of compound 3, it did not substantially decrease the amount of cathepsin B in non-tumoral cell lines (p < 0.001), presenting higher K i (K i = 3.29 ± 0.40 nM, TKPTS cells and K i = 4.55 ± 0.81nM, FL83B cells) ( Figure 3A). Figure 3B shows the results obtained in both tumour cell lines (ACHN and Hepa-1c1c7 cells) and two non-tumour cell lines (TKPTS and FL83B cells) in relation to cathepsin B's activity, which was higher against tumour cells than non-tumour ones (p < 0.001).
Cathepsin B is a 30-kDa, bilobal protein with an active site and substrate-binding cleft located in the interface of a left (L-) and a right (R-) domain with three and two binding sites located in the loops, categorised as S3, S1, and S2', and S1' and S2, respectively. Access of the substrate into the active site is controlled by the occluding loop, which consists of an 18-residue-long insertion (Pro 107-Asp 124) [27].
Based on the structure of this enzyme, it is known that the catalytic triad of cathepsin B is formed by a cysteine, a histidine, and an aspartic acid. The Cys29 and His199 residues interact, catalysing the cleavage of the peptide bond. Therefore, cathepsin B inhibitors must have an electrophilic character capable of reacting with the thiol group of the Cys29 residue [17]. Examples include synthetic 'warheads' such as aldehydes, disulphides, vinyl sulfones, and halomethyl ketones [28]. In our work, we used the compound CA 074 as a positive control, whose mechanism of action is performed through its binding to the S' subsites in a similar direction as the substrate [17].
All the compounds (one isolated and two synthesised) share the essential premise of maintaining a main oxadiazine scaffold, which is structurally similar to thiadiazols. In this sense, a cleavage of the N-O bond of oxadiazine by the thiolate group of cathepsin B could produce the opening of the ring and proteolytic enzymatic inhibition [17].
However, in the case of compound 2, to improve the membrane permeability, an ester-type derivative was introduced. Nevertheless, the activity results obtained were worse (loss of selectivity) compared to compound 1 and the positive control. These data agree with the consulted bibliography since the esterification of CA 074 (Prodrug) improves its permeability through cell membranes but decreases its activity and selectivity [29].
Finally, biological assays showed that compound 3 was the most active of the three compounds. Analysing its structure, we observed that the presence of a free carboxylate group improved the selectivity for cysteine proteases in comparison to other nucleophilic biological sulphides [17]. In addition, the presence of this free carboxylate group allowed interactions with the His111 residue of the occlusion loop, improving the selectivity of the compounds [30].
Although previous in vitro studies suggest that compound 3 may be a potential cathepsin B inhibitor, additional molecular coupling analyses are required for a better understanding of the molecule-enzyme interaction.

Molecular Docking of the Active Compounds
To confirm the interactions of compounds (1-3) and of the positive control (CA 074) with the cathepsin B protease (PDB: 1QDQ), in silico docking studies were performed using GOLD software [31], which deploys a genetic algorithm (GA) to generate different ligand conformations and defines the binding site using the reference inhibitor CA 074, using a cut-off distance of 0.8 Å.
On the other hand, docking solutions were evaluated with the ChemPLP scoring function [32] and the GoldScore re-scoring function [33]. Subsequently, the best-scored poses of the most populated groups were selected for further analysis, and the ones that did not include interactions with essential residues for the enzymatic activity were discarded. It is important to highlight that while crystallography provides detailed information about ligand binding, docking provides guidance on what that binding might look like and should not be taken as definitive since there may be variations, for example, in hydrophobic interactions depending on the position of the alkyl tails. Therefore, a more detailed interpretation must be supported by further binding studies (Figure 4). poses of the most populated groups were selected for further analysis, and the ones that did not include interactions with essential residues for the enzymatic activity were discarded. It is important to highlight that while crystallography provides detailed information about ligand binding, docking provides guidance on what that binding might look like and should not be taken as definitive since there may be variations, for example, in hydrophobic interactions depending on the position of the alkyl tails. Therefore, a more detailed interpretation must be supported by further binding studies (Figure 4). Analysis of the interactions between compound 3 and cathepsin B revealed the formation of hydrogen bonds between Cys29A, Gln29A, and His110 and the molecule, in addition to hydrophobic interactions with other residues such as Ala200A. This cysteine-29 residue is part of the catalytic triad responsible for enzyme activity [34] and, typically, inhibitors must have an electrophilic character capable of reacting with cysteine residue 29 [17]. This suggests that compound 3 binds similarly to the reference inhibitor CA 074 and other compounds described previously [27,35]. In addition, the carboxylate group present in compound 3 establishes interactions with the His110A residue, which has also been described as an important residue for the binding of ligands to cathepsin B [36] (Figure 5). Analysis of the interactions between compound 3 and cathepsin B revealed the formation of hydrogen bonds between Cys29A, Gln29A, and His110 and the molecule, in addition to hydrophobic interactions with other residues such as Ala200A. This cysteine-29 residue is part of the catalytic triad responsible for enzyme activity [34] and, typically, inhibitors must have an electrophilic character capable of reacting with cysteine residue 29 [17]. This suggests that compound 3 binds similarly to the reference inhibitor CA 074 and other compounds described previously [27,35]. In addition, the carboxylate group present in compound 3 establishes interactions with the His110A residue, which has also been described as an important residue for the binding of ligands to cathepsin B [36] (Figure 5). Finally, a correlation was made between the coupling scores (obtained from 300 different poses and evaluated through the affinity (PLP score) and rescore (GOLD) functions) and the inhibition constants (Ki) of the samples (Figure 6). Finally, a correlation was made between the coupling scores (obtained from 300 different poses and evaluated through the affinity (PLP score) and rescore (GOLD) functions) and the inhibition constants (Ki) of the samples (Figure 6). Finally, a correlation was made between the coupling scores (obtained from 300 different poses and evaluated through the affinity (PLP score) and rescore (GOLD) functions) and the inhibition constants (Ki) of the samples (Figure 6). Using the box and whisker plot, we can conclude that there is not a perfect correlation between the coupling scores and the experimental Ki. This tells us that the exact binding mode of each compound to cathepsin B needs to be further investigated through structural biology experiments. Using the box and whisker plot, we can conclude that there is not a perfect correlation between the coupling scores and the experimental Ki. This tells us that the exact binding mode of each compound to cathepsin B needs to be further investigated through structural biology experiments.

Acute Toxicity In Vivo
Based on the preliminary positive results of in vitro cytotoxicity, we decided to carry out in vivo toxicological studies at the same dose as the positive control, CA 074 (4 mg/kg), to determine whether the compounds showed unwanted effects on mice. For this, mice were divided into five groups, which were given different treatments based on the presence of the positive control and compounds 1, 2, or 3 (groups two, three, four, and five, respectively). After the third day of treatment, mice in groups two (4 mg/kg i.p., CA 074), three (4 mg/kg i.p., compound 1), and five (4 mg/kg i.p., compound 3) were active and showed no apparent weight loss. However, in group four (4 mg/kg i.p., compound 2), weight loss and signs of disease (bowed backs, loss of hair shine, and slow movements) were observed.
It was not possible to administer higher doses because this study aimed to determine whether the compounds were toxic at the same concentration as the positive control (CA 074), so the LD 50 values were not determined and the degree of involvement of the mice was measured subjectively on a gradual scale from 0 to 10; where higher values represent an increase in physical harm observed in animals, including death.
After concluding the week of treatment, major signs of toxicity were not detected (macroscopic observation). The animals were sacrificed, and histological brain, intestine, heart, liver, kidney, and pancreas sections were performed. Histological analysis of mice in all groups showed no inflammation, necrosis, fibrosis, or hyperplasia signs. Even group four, where mice showed signs of disease on the third day, revealed no signs of toxicity or damaged organs (Figure 7). an increase in physical harm observed in animals, including death.
After concluding the week of treatment, major signs of toxicity were not detected (macroscopic observation). The animals were sacrificed, and histological brain, intestine, heart, liver, kidney, and pancreas sections were performed. Histological analysis of mice in all groups showed no inflammation, necrosis, fibrosis, or hyperplasia signs. Even group four, where mice showed signs of disease on the third day, revealed no signs of toxicity or damaged organs (Figure 7).

General Experimental Procedures
First-grade organic solvents, purchased from Sigma-Aldrich, were used for the isolation of the compounds. Thin-layer chromatography (TLC) was performed using Merck Silica gel 60-F254 plates. The different chromatoplates were visualised by UV absorbance (254 nm) and stained with phosphomolybdic acid followed by heating. A silica gel (40-63 μm and 20-45 μm, Merck) chromatography column was prepared using the indicated eluent in accordance with standard techniques.
NMR experiments were performed with Bruker Avance DRX 300 and 500 spectrometers operating at 300 MHz/500 MHz ( 1 H) or 75 MHz/175 MHz ( 13 C). Chloroform-d1 and DMSO-d6 were used as deuterated solvents. The spectra were calibrated by assigning the peaks of the residual solvents. HREIMS spectra were obtained through electronic impact techniques (EI+), at 70 e.V., and electrospray (ESI+), using a QSTAR XL quadrupole TOF mass spectrometer. MS samples were prepared in MeOH.

General Experimental Procedures
First-grade organic solvents, purchased from Sigma-Aldrich, were used for the isolation of the compounds. Thin-layer chromatography (TLC) was performed using Merck Silica gel 60-F 254 plates. The different chromatoplates were visualised by UV absorbance (254 nm) and stained with phosphomolybdic acid followed by heating. A silica gel (40-63 µm and 20-45 µm, Merck) chromatography column was prepared using the indicated eluent in accordance with standard techniques.
NMR experiments were performed with Bruker Avance DRX 300 and 500 spectrometers operating at 300 MHz/500 MHz ( 1 H) or 75 MHz/175 MHz ( 13 C). Chloroform-d 1 and DMSO-d 6 were used as deuterated solvents. The spectra were calibrated by assigning the peaks of the residual solvents. HREIMS spectra were obtained through electronic impact techniques (EI+), at 70 e.V., and electrospray (ESI+), using a QSTAR XL quadrupole TOF mass spectrometer. MS samples were prepared in MeOH.

Biological Material
A sample of the cyanobacterium Nostoc sp. (BEA0798B) was collected from Garajonay National Park, La Gomera, Gran Canarias (Canary Islands, 28 • 07 40.7"N and 17 • 14 12.6"W, Spain). The 16S ribosomal RNA gene sequence was deposited in GenBank (accession No. OQ379060.1). The sample was cultured in 2 L flasks with BG-110 medium under axenic conditions by VALORALIA I MÁS D, SL. Cultures were grown with an air bubbling at 23 ± 1 • C and ambient light for 21 days. After this time, the biomass generated was separated by centrifugation and, finally, lyophilised.

Extraction and Isolation
The lyophilised biomass (15.1 g) was extracted by repeated maceration with CH 2 Cl 2 / MeOH (1:1, 3 × 1 L) to obtain a dichloromethane/methanol extract (3.17 g). The biological activity (apoptotic capacity and cathepsin B inhibition) of this extract was measured against tumour and non-tumour cell lines.

Assay for Activity of Cathepsin B
Cathepsin B activity in the cell medium was evaluated by using a fluorogenic substrate: Abz-Gly-Ile-Val-Arg~Ala-Lys(Dnp)-OH. This peptide substrate was purchased from Bachem AG (Bubendorf, Switzerland). The reactions were performed at 40 • C for 10 min [38]. Stock solutions of inhibitory compounds were prepared in DMSO (in a concentration of 10 mM). From these stock solutions, increasing concentrations of the inhibitory compounds (1 and 100 nM) were prepared. L-3-trans-(propylcarbamyl)-oxirane-2-carbonyl)-L-isoleucyl-L-proline (CA 074 ≥ 99%Sigma-Aldrich, CAS number 134448-10-5) was used as a positive control [39]. Fluorescence was measured with excitation/emission wavelength values of 320 and 420 nm, respectively, using the Cary Eclipse fluorescence spectrophotometer (Agilent Technologies, Melbourne, Australia).

In Vivo Toxicity
Adult male C57BL/6 mice weighing 25-35 g were purchased from the Centre for Laboratory Animals of the Faculty of Medicine from the University Autónoma of Madrid (Spain) and maintained under controlled temperature conditions (22 ± 2 • C), with a constant 12 h light-dark cycle and with access to food and water ad libitum. The experiments reported in this study were performed in accordance with the guidelines of the EU Directive Five groups (five mice per group) were selected for this study. Group one served as a negative control and received a normal saline solution at a dose of 5 mL/kg (i.p.), group two served as a positive control and received compound CA 074 at a concentration of 4 mg/kg, i.p. [40], group three were treated with the natural compound (1) (4 mg/kg, i.p.), group four were treated with compound 2 (4 mg/kg, i.p.), and group five were treated with compound 3 (4 mg/kg, i.p.). The concentration of 4 mg/kg of the natural compound and the two synthesised compounds were chosen based on the concentration reported for compound CA 074. This is because our study aimed to compare the toxicity of our compounds with compound CA 074, which is a cathepsin B inhibitor that reduces tumour cell metastasis.
All animals were kept under observation for 7 days, during which their physical aspect, behaviour, and the number of deaths were recorded daily. The negative control (normal saline solution) received only the same quantity of the vehicle (distilled water). These tests were carried out in triplicate. All the surviving animals were euthanised at the end of the study, and their vital organs were individually observed by necropsy and checked for overt pathology.

Statistical Analysis
All analyses were performed using version 9.0.0 of GraphPad Prism Software LLC from 1994-2020 (www.graphpad.com, accessed on 2 January 2023). Using non-linear regression, the values of 50% cytotoxic concentration (CC 50 ) and 50% inhibitory concentration (IC 50 ) were determined. In addition, a one-way ANOVA statistical analysis was performed to assess whether the differences between the values were statistically significant (p < 0.05; p < 0.001, Tukey's multiple comparisons test). All experiments were performed in triplicate.

Computational Methods
In silico docking studies for compounds 1-3 and the positive control (CA 074) were carried out using GOLD software [31] (Hermes 2022.2.0 (Build 353591)). The protocol was validated by removing the co-crystallised ligand of the receptor, and by re-docking it into the active site (RMSD was used for evaluation of results). For energy-minimised ligand docking, 300 GA runs, and 125,000 operations were used.

Supplementary Materials:
The following supporting information can be downloaded at: https: //www.mdpi.com/article/10.3390/md21050284/s1, NMR and HRESIMS spectra of the synthetic oxadiazines analysed in this study are provided as supporting information (Figures S1-S27).